It doesn't matter at all
Unless you have a strange user case with a 2 step flow (mail first / password second), the validation will come after the user completes both fields and submits the form. You can validate with autocomplete, but that would be a security issue, so I assume you will do it like everyone else.
About your disadvantages:
- Check email only: the fact that the user is banned will be known by anyone who uses their email to log in.
Why worry? Banned someone and worried that someone is trying to illegally use someone else's account? Neither of them is your user, they are both breaking your rules, why do you care about them and their feelings?
- Verify email and password: The user will forget the account and it is possible for someone to access the website and test a password to make sure it is used by the user.
I'm not sure I have understood this correctly: on the one hand, it would be the same case (someone else trying to break your security measures using the credentials of an already banned person). Otherwise the only reason for this would be a super weird and extreme extreme case, like 1 in 1 million. But again: I'm not sure what you mean exactly.
In short
If the banned user tries to enter the site, just show him a notice after he tried. It's a proven pattern, no reason to change it.